Visualize what a processor feeds a vision–text model

Models

Output